Efficient Urban Mobility with AI Traffic Solutions

Urban traffic congestion is a major challenge in rapidly growing cities, causing delays, increased fuel consumption, and environmental pollution. This study addressed the problem of inefficient urban mobility in Port Harcourt, where peak-hour traffic led to significant congestion at major junctions, including Garrison Junction and Rumuokoro Junction. To solve this problem, a Naïve Bayes-based traffic prediction system was developed to forecast congestion patterns using day, time, and location as features. A synthetic dataset was generated for model training and testing, and predictions were made for different hours of the day. The results showed that the model achieved an overall accuracy of 85%, with precision of 0.83, recall of 0.87, and F1-score of 0.85. Temporal analysis revealed peak congestion during morning (7:00–9:00) and evening (16:00–20:00) periods, while mid-day periods exhibited lower congestion (average 0.39). Spatial analysis showed that Garrison Junction had higher congestion (0.62) compared to Rumuokoro Junction (0.44), reflecting location-specific traffic patterns. The system successfully captured both temporal and spatial congestion trends, demonstrating its effectiveness in supporting intelligent traffic management and efficient urban mobility. The system was implemented using Python, with pandas, numpy, scikit-learn, and matplotlib for data processing, modeling, and visualization. These results highlighted the potential of AI traffic solutions to enhance urban mobility planning and informed decision-making for city authorities.